summary.ame {FLAME}

Summarize the output of FLAME or DAME

Description

These methods create and print objects of class summary.ame containing information on the numbers of units matched by the AME algorithm, matched groups formed, and, if applicable, average treatment effects.

Usage

## S3 method for class 'ame'
summary(object, ...)

## S3 method for class 'summary.ame'
print(x, digits = 3, ...)

Arguments

object

An object of class ame, returned by a call to FLAME or DAME

...

Additional arguments to be passed on to other methods.

x

An object of class summary.ame, returned by a call to summary.ame

digits

Number of significant digits for printing the average treatment effect estimates and their variances.

Details

The average treatment effect (ATE) is estimated as the average CATE estimate across all matched units in the data, while the average treatment effect on the treated (ATT) and average treatment effect on controls (ATC) average only across matched treated or matched control units, respectively. Variances of these estimates are computed as in Abadie, Drukker, Herr, and Imbens (The Stata Journal, 2004) assuming constant treatment effect and homoscedasticity. Note that the implemented estimator is not =asymptotically normal and so in particular, asymptotically valid confidence intervals or hypothesis tests cannot be conducted on its basis. In the future, the estimation procedure will be changed to employ the nonparametric regression bias adjustment estimator of Abadie and Imbens 2011 which is asymptotically normal.

Value

A list of type summary.ame with the following entries:

MG

A list with the number and median size of matched groups formed. Additionally, two of the highest quality matched groups formed. Quality is determined first by number of covariates matched on and second by matched group size.

n_matches

A matrix detailing the number of treated and control units matched.

TEs

If the matching data had a continuous outcome, estimates of the ATE, ATT, and ATC and the corresponding variance of the estimates.